The gymnosperm Ginkgo biloba, of the Conopsida class, Ginkgoales order, and Ginkgoaceae family, originated in Eastern China approximately 150 million years ago and is the sole living representative of its order (Schwarz and Arigoni, 1999; Benson, L., 1957; Chaw, et al., 2000; Bowe, et al., 2000). This hardy tree, termed a “living fossil” by Charles Darwin, is well-known for its ability to withstand harsh climate conditions and resist insect infestation (Major, R. T., 1967). The apparent lack of change over millions of years is presumably due to its long time span between generations; reproduction begins after 20 years of age and continues to 1000 years of age.
G. biloba is renowned as a potent herbal therapeutic that aids in the revascularization of ischemic tissue through improved microcirculation. G. biloba leaf extracts have been used for centuries to treat cerebrovascular and cardiovascular diseases, dementia, tinnitus, arthritis, and vertigo (Itil, et al., 1995; Briskin, D. P., 2000). These beneficial pharmacological effects have been attributed, in part, to the ginkgolides, a unique series of diterpene molecules which are highly specific platelet-activating factor (PAF) receptor antagonists (Hosford et al., 1990). Generation of PAF occurs during anaphylaxis or shock and leads to bronchoconstriction, contraction of smooth muscle, and reduced coronary blood flow, which are often fatal. The isomer known as ginkgolide B demonstrates the highest activity of the diterpenes and antagonizes all known PAF-induced membrane events. Furthermore, the American Medical Association recently endorsed the Chinese herb as a viable alternative to traditional approaches in the treatment of Alzheimer's disease. Recent studies report that the extract delayed the progression of dementia in approximately one third of the patients studied (Le Bars et al., 1997).
Ginkgolides were first isolated from the roots of the Ginkgo tree by Furukawa (1932) and later characterized by K. Nakanishi (1967) and Sakabe (1967); the elucidated structures were named Ginkgolides A, B, C, J, and M. In 1967, K. Okabe also established the presence of the ginkgolides in the leaves of the Ginkgo tree. Ginkgolides are biosynthesized from geranylgeranyl diphosphate, the universal diterpene precursor. These molecules contain a caged trilactone structure and display a rare tert-butyl group. Analogs are distinguished by the number and location of hydroxyl group substituents. Recently, the ginkgolides and bilobalide (a pentanorditerpenoid by-product of ginkgolide biosynthesis) were determined to have significant antifeedant activities toward insect larvae (Schwarz, M., 1994; Matsumoto, et al., 1987).
Geranylgeranyl diphosphate (GGDP) (Schwarz and Arigoni, 1999) employed in ginkgolide biosynthesis is derived from isopentenyl diphosphate formed via the deoxyxylulose pathway. The proposed biosynthesis of the ginkgolides is initiated by protonation of GGDP to give labdadienyl diphosphate. Ionization of the allylic diphosphate moiety followed by a 1,4-hydrogen shift, methyl migration, and deprotonation yields levopimaradiene (Schwarz and Arigoni, 1999). The proposed intramolecular hydrogen shift was also observed in the biosynthesis of Abies grandis abietadiene synthase (AgAS) (Ravn et al., 1998; Ravn et al., 2000). Oxidation of ring C produces abietatriene, which is then transported from the plastid to the cytoplasm. The aromatic hydrocarbon undergoes further transformation in the endoplasmic reticulum by cytochrome P450-dependent monooxygenases to produce the ginkgolides (Schwarz and Arigoni, 1999) (FIG. 1).
Metabolic regulation studies of diterpene production in G. biloba seedlings indicate that ginkgolides are produced in the roots and are subsequently translocated to the leaves. Furthermore, diterpene hydrocarbon precursors were found exclusively in the roots and included levopimaradiene, palustradiene, abietadiene, pimaradiene, and abietatriene. Addition of cytochrome P450-dependent oxygenase inhibitors to the roots of seedlings resulted in full inhibition of oxygenation reactions along the pathway to the diterpenes. Abietatriene, the sole diterpene hydrocarbon obtained, was identified as the immediate precursor to the ginkgolides (Cartayrade et al., 1997; Neau et al., 1997).
Presently, commercial development of the ginkgolides as therapeutic agents has been hampered. Because these diterpenoids contain up to 12 stereocenters, 4 contiguous quaternary carbons, and 3 oxacyclic rings fused to 2 spiro carbocyclic rings, they present a formidable synthetic challenge. In spite of the topological and stereochemical complexity inherent to the ginkgolides, total syntheses of these unusually challenging targets have been achieved. In 1988, the first synthesis of (±)-ginkgolide A (38 steps, <1% overall yield) and (±)-ginkgolide B (35 steps, <1% overall yield) were reported (Corey and Ghosh, 1988; Corey et al., 1988). Furthermore, ginkgolide B was converted to ginkgolide A in 6 steps and approximately 50% yield. More recently, (±)-ginkgolide B was synthesized in 26 steps and 3% total yield (Crimmins et al., 1999). Although strategically impressive, these demanding routes require multiple transformations resulting in low yields that ultimately preclude commercial-scale production of the ginkgolides.
Current commercial ginkgolide production relies exclusively on extraction from Ginkgo trees, which accumulate low levels of the compound. In addition, the demand for this medicinal plant has increased at a rate of 26% per annum with 2,000 tons harvested annually (Masood, E., 1997) G. biloba plantations serve as the major source of the herbal extract and provide an average 1 to 7 mg/g dry weight ginkgolide from young trees (Balz, et al., 1999) In an effort to increase diterpenoid content, G. biloba seedlings, plants, and trees were treated with metabolic inhibitors that target key branchpoints in isoprenoid biosynthesis downstream of GGPP synthesis (Huh, et al., 1993) Presumably, inhibiting GGPP depleting pathways would increase the available concentration of GGPP, the natural diterpene substrate. Variable results were obtained with cycloartenol synthase inhibitors, ancymidol and AMO-1618. In contrast, application of fluridone (an inhibitor of carotenoid biosynthesis that blocks phytoene desaturation) yielded up to 78% more ginkgolides.
Extraction of the ginkgolides from G. biloba is known. U.S. Pat. No. 5,399,348 refers to a method for preparation of Ginkgo biloba extract in which the alkylphenol compounds are separated not by using chlorinated aliphatic hydrocarbon, but through a process of precipitation, filtration and multi step liquid-liquid-extractions. U.S. Pat. Nos. 5,399,348; 5,322,688; 5,389,370; 5,389,370; 5,637,302; 5,512,286; 5,399,348; and 5,389,370 are all directed to various methods of preparing a desired Ginkgo biloba extract. U.S. Pat. Nos. 5,241,084 and 5,599,950 are directed to methods to convert ginkgolide C to ginkgolide B.
Seeking an alternative, non-synthetic approach to ginkgolide production, a method to clone and functionally express genes involved in their biosynthesis was considered. In 1971, the isoprenoid nature of the ginkgolides was precariously, yet correctly, established using 2-14C MVA incorporation experiments conducted with G. biloba seedlings. Moreover, the researchers proposed that the unique tert-butyl group arose from S-adenosyl methionine (Nakanishi, et al., 1971). However, a revised biogenetic scheme was put forth as a result of NMR product analyses of isotopically labeled precursors incubated with G. biloba embryos (Schwarz, et al., 1999). During the course of these extensive studies, a dichotomy was observed concerning the biosynthesis of IPP by G. biloba. Specifically, formation of isopentenyl pyrophosphate (IPP), an isoprene unit possessing a diphosphate moiety, proceeds via the classical MVA pathway in the synthesis of sitosterol, but in the plastids, the deoxyxylulose-5-phosphate (DXP) pathway synthesizes GGPP. Presumably, segregation between the two pathways is due to compartmentalization of the plant cell. IPP responsible for sitosterol formation is restricted to the cytoplasm, and IPP incorporated into ginkgolides originates in the chloroplasts.
G. biloba cultures were first established in 1971; however, HPLC analysis failed to detect ginkgolides (Nakanishi, et al., 1971). Two decades later, ginkgolides A and B were detected in undifferentiated cell cultures (<20 ng/g dry weight), albeit by a factor of 106 less than that obtained from leaves of mature trees (Carrier, et al., 1991; Chauret, et al., 1991). Increased ginkgolide content was demonstrated in primary callus and cell suspension cultures (˜26% and 47% relative to leaves of mature trees, respectively) were unable to be maintained in secondary cultures (Huh, et al., 1993). Currently, high yield production of the ginkgolides by in vitro cultures of undifferentiated cells has not been achieved (Balz et al., 1999). Transgenic cells were obtained from putative transformed G. biloba embryos but ginkgolide concentration was <400 μg/g dry tissue culture (Laurain, et al., 1997). Recently, Dupréet al. (2000) reported a reproducible transformation protocol of G. biloba by Agrobacterium tumefaciens; however, ginkgolide levels of the transformed cells have not been disclosed.
There are examples in the art in which heterologous diterpene synthases are introduced into and expressed in organisms such as Escherichia coli, particularly for the purpose of characterizing activity of a soluble form of the enzyme in the absence of any plastidial targeting sequence (Hill et al., 1996; Peters et al., 2000; Williams et al., 2000). However, the novel levopimaradiene synthase of the present invention provides a solution to a need in the art for methods and compositions to quickly produce large amounts of substantially pure ginkgolides in a cost-effective manner, particularly in an organism capable of a high-yield ginkgolide-producing system.